Panos Papaspyridakos Digital versus conventional implant

German O. Gallucci
Chun-Jung Chen
impressions for edentulous patients:
Stijn Hanssen accuracy outcomes
Ignace Naert
Bart Vandenberghe

Authors’ affiliations:
Panos Papaspyridakos, Division of Postgraduate
Prosthodontics, Tufts University School of Dental
Medicine, Boston, MA, USA
Department of Prosthodontics, School of Dentistry,
National and Kapodistrian University of Athens,
Athens, Greece
German O. Gallucci, Division of Regenerative and
Implant Sciences, Harvard School of Dental
Medicine, Boston, MA, USA
Chun-Jung Chen, Department of Dentistry, Chi
Mei Medical Center, Tainan, Taiwan
Stijn Hanssen, Layerwise, Leuven, Belgium
Ignace Naert, Bart Vandenberghe, Prosthetics
Section, Department of Oral health Sciences,
Catholic University of Leuven, Leuven, Belgium
Corresponding author:
Dr. Panos Papaspyridakos
Division of Postgraduate Prosthodontics
Tufts University School of Dental Medicine
1 Kneeland Street, Boston 02111 MA
USA
Tel: +16176366828
Fax: +16176360469
e-mail: panpapaspyridakos@gmail.com
Key words: accuracy, dental implants, digital impressions, edentulous, implant impressions,
impression techniques
Abstract
Purpose: To compare the accuracy of digital and conventional impression techniques for
completely edentulous patients and to determine the effect of different variables on the accuracy
outcomes.
Materials and methods: A stone cast of an edentulous mandible with five implants was fabricated
to serve as master cast (control) for both implant- and abutment-level impressions. Digital
impressions (n = 10) were taken with an intraoral optical scanner (TRIOS, 3shape, Denmark) after
connecting polymer scan bodies. For the conventional polyether impressions of the master cast, a
splinted and a non-splinted technique were used for implant-level and abutment-level impressions
(4 cast groups, n = 10 each). Master casts and conventional impression casts were digitized with an
extraoral high-resolution scanner (IScan D103i, Imetric, Courgenay, Switzerland) to obtain digital
volumes. Standard tessellation language (STL) datasets from the five groups of digital and
conventional impressions were superimposed with the STL dataset from the master cast to assess
the 3D (global) deviations. To compare the master cast with digital and conventional impressions
at the implant level, analysis of variance (ANOVA) and Scheffe’s post hoc test was used, while
Wilcoxon’s rank-sum test was used for testing the difference between abutment-level conventional
impressions.
Results: Significant 3D deviations (P < 0.001) were found between Group II (non-splinted, implant
level) and control. No significant differences were found between Groups I (splinted, implant
level), III (digital, implant level), IV (splinted, abutment level), and V (non-splinted, abutment level)
compared with the control. Implant angulation up to 15° did not affect the 3D accuracy of implant
impressions (P > 0.001).
Conclusion: Digital implant impressions are as accurate as conventional implant impressions. The
splinted, implant-level impression technique is more accurate than the non-splinted one for
completely edentulous patients, whereas there was no difference in the accuracy at the abutment
level. The implant angulation up to 15° did not affect the accuracy of implant impressions.

Passive fit of implant-fixed complete dental
prosthesis (IFCDP) depends on the accuracy
of the implant cast, which is directly depen-
dent on the accuracy of the impression tech-
nique (Jemt & Hjalmarsson 2012;
Papaspyridakos & Lal 2013). There are sev-
eral clinical and laboratory variables that
Date: affect the accuracy of an implant cast,
Accepted 15 January 2015
namely impression and pouring techniques,
To cite this article: impression material and die stone properties,
Papaspyridakos P, Gallucci GO, Chen C-J, Hanssen S, Naert I,
Vandenberghe B. Digital versus conventional implant machining tolerance of prosthetic compo-
impressions for edentulous patients: accuracy outcomes. nents, and implant angulation and depth (Ma
Clin. Oral Impl. Res. 00, 2015, 1–8
doi: 10.1111/clr.12567 et al. 1997; Papaspyridakos et al. 2014a).
The first and most significant step is the
impression procedure. Different implant
impression techniques have been used to
generate a definitive cast that will ensure
the accurate clinical fit of IFCDPs. A recent
systematic review on the accuracy of
implant impressions showed that the
splinted technique is superior to the non-
splinted option for both partially and com-
pletely edentulous patients (Papaspyridakos
et al. 2014a). The necessity for splinting the
impression copings has been advocated in
several studies, while others have shown no

© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd 1
Papaspyridakos et al
Digital versus conventional implant impressions
difference (Papaspyridakos et al. 2012). It has
also been reported that open-tray techniques
are superior to the closed-tray version for com-
pletely edentulous patients, but no difference
was identified for partially edentulous patients
(Papaspyridakos et al. 2014a). The diverse
results from some previous in vitro studies
may be partially explained by the machining
tolerance of components, by the differences in
methods for accuracy measurements, and by
improvements in dental materials.
The effect of implant angulation and con-
nection type on the accuracy of implant casts
generated with various impression tech-
niques for edentulous jaws has not been fully
investigated yet. For completely edentulous
patients, six in vitro and three clinical stud-
ies reported on accuracy outcomes with
angulated implants (Aguilar et al. 2010;
Papaspyridakos et al. 2011, 2012; Mpikos
et al. 2012; Ongul et al. 2012; Stimmelmayr
et al. 2012a,b; Akalin et al. 2013; Gimenez
et al. 2015). The three clinical studies did not
focus on the details of implant angulation
but reported that the splinted technique was
clinically more accurate than non-splinted or
closed-tray techniques when angulated
implants were involved (Papaspyridakos et al.
2011, 2012; Stimmelmayr et al. 2012b),
whereas the six in vitro studies showed
mixed results. Hence, there is insufficient
guidance for choosing impression techniques
for different implant angulations in com-
pletely edentulous cases. Additionally, the
effect of implant connection type (internal
connection vs external connection) on the
accuracy of implant impressions for fully
edentulous jaws has not been investigated
sufficiently.
Digital implant dentistry has transformed
the relationship between dentist and dental
laboratory. As a part of this trend, digital
impressions have been a significant contribu-
tor to this changing relationship. Digital
impression scanners eliminate tray selection,
dispensing and setting of impression materi-
als, disinfection, and impression shipping to
the laboratory, while patient comfort may be
an additional advantage (Christensen 2009;
Ender & Mehl 2013; Patzelt et al. 2014). In-
traoral scanners may increase efficiency,
because it is possible to digitally send a digi-
tal impression to the laboratory, rather than
sending a conventional impression via regu-
lar mail. The digital impression file can be
stored electronically, which eliminates space
management issues, supports a paper-free
practice, and contributes to efficient record
keeping. Limitations pertain to the additional
cost of purchasing an intraoral scanner and
2 | Clin. Oral Impl. Res. 0, 2015 / 1–8
the learning curve for adjusting to the new
treatment modality.
A growing number of edentulous patients
are seeking implant prosthodontic treatment.
The application of computer-guided surgery
and CAD/CAM technology in implant pros-
thodontics has aided in simplifying a number
of treatment steps (Papaspyridakos et al.
2014a, 2014b). No data exist on the accuracy
of digital implant impressions for completely
edentulous jaws. Research on digital implant
impressions is limited to a few case reports
(Lin et al. 2013, 2014; Moreno et al. 2013).
Therefore, a study assessing the performance
and accuracy of digital impressions compared
to conventional impressions for completely
edentulous patients would contribute to
clinically validate this new cutting-edge
technology.
The purposes of this study were (i) to test
whether or not digital implant impressions
are more accurate than conventional implant
impressions in completely edentulous
patients and (ii) to test whether or not the
implant angulation or the implant connec-
tion type (internal connection vs abutment-
level connection) affects the accuracy of
implant impressions and casts of completely
edentulous patients. The null hypothesis of
this investigation was that digital impres-
sions exhibit similar accuracy as conven-
tional implant impressions.
Materials and methods
A mandibular cast with five interforaminal
internal connection implants (Bone Level
RC, Straumann, Basel, Switzerland) was fab-
ricated to simulate a common clinical situa-
tion. The median three implants were
parallel to each other, whereas the distal left
implant had an angulation of 10° and the dis-
tal right of 15°. This cast was fabricated in
clear acrylic resin in a specialized facility
(Model Plus Inc, Grayslake, Il).
Master cast fabrication
As the clear acrylic resin cast could not be
digitally scanned and digitized, a stone mas-
ter cast was fabricated to serve as control.
One screw-retained metal framework (laser-
welded titanium bars to titanium abutments)
was fabricated on this clear acrylic resin cast
at the abutment level (Multi-Base RC; Strau-
mann). A pickup impression of the master
implant framework was taken with polyether
impression material (3M ESPE; Impregum,
St. Paul, MN, USA). Implant analogs were
then connected to the framework inside the
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
pickup impression, which was poured with
low expansion (0.09%) type IV die stone
(Silky Rock; Whipmix Corp, Louisville, KY,
USA) 2 h after impression taking (Del’Acqua
et al. 2008). The stone was mixed under vac-
uum with distilled water, and an initial pour
of stone up to the middle of the analogs was
carried out. After 30 min, the second pour of
vacuum-mixed die stone was added. The
stone cast was allowed to set for 1 h, as per
manufacturer’s recommendation, before sepa-
rating it from the impressions, trimming, and
finishing.
Once the master cast (control) was com-
pleted, a custom tray was fabricated after
four fiducial mark stops were made on the
master cast to standardize custom tray posi-
tioning during open-tray impression taking
(Fig. 1a,b). During custom tray fabrication,
two layers of baseplate wax (NeoWax; Dents-
ply Inc, York, PA, USA) were applied to pro-
vide 3 mm of space relief for the impression
material. The custom tray, with five holes to
accommodate the impression coping guide
pins and four mark stops, was fabricated with
visible light-curing acrylic resin (Triad Tru-
Tray VLC; Dentsply Inc, York, PA, USA). A
box for pouring the impression with dental
stone was made with addition reaction sili-
cone (Exaflex putty; GC America Inc, Alsip,
Il) to create a silicone matrix. This matrix
was used for pouring all the impressions,
allowing standardization of the shape of the
stone casts and for the amount of dental
stone used for the pouring.
(a)
(b)
Fig. 1. Master cast (control) (a) implant-level (b)
abutment-level.

Conventional implant impression procedures
Prior to the impression, tray adhesive (3M
ESPE; Impregum) was applied thinly and
evenly into the intaglio surface of the custom
tray. The tray adhesive was allowed to dry
for 15 min before the impression was made.
Polyether impression material (Impregum)
was used for all conventional implant
impressions. The impressions were made in a
controlled-temperature environment (25°C 
2°C) with a relative humidity of 50%. The
mark stops on the master cast served as an
index for precise positioning of the custom
tray every time. The custom tray was always
seated with light finger pressure until the
mark stops contacted their respective areas
on the master cast. The polyether impression
material was allowed to polymerize for
12 min. Four different techniques were used
for impression taking: (i) implant-level,
splinted technique with visible light-cured
resin (Triad gel; Dentsply Inc, York, PA), (ii)
implant-level, non-splinted impression tech-
nique, (iii) abutment-level, splinted technique
with visible light-cured resin (Triad gel;
Dentsply Inc, York, PA, USA), and (iv) abut-
ment-level, non-splinted impression tech-
nique. Ten implant impressions with
polyether material (Impregum) were taken to
fabricate 10 implant casts for each technique
(n = 10). Standardized pressure was applied
over each custom tray while setting, by the
help of 1 kg weight.
Splinting materials and technique
The splinting material used was urethane di-
methacrylate-based visible light-cured resin
(Triad gel; Dentsply Inc, York, PA, USA). Pre-
fabricated resin bars had been made by filling
drinking straws with resin (Triad gel) followed
by light curing, thus creating resin bars of
standardized thickness and shape. The resin
bars were stored for 24 h and then used to
splint the impression copings together with
the aid of additional light-cured resin (Triad
gel). The splints were sectioned with a disk
and re-connected with minimal amount of
visible light-cured resin to compensate for
polymerization shrinkage. The splint was left
untouched for 5 min (Papaspyridakos et al.
2011, 2012). For the non-splinted technique,
the impression copings were not splinted to
each other. For the abutment-level impres-
sions, straight multiunit abutments (Multi-
Base RC, Straumann) of the same height were
torqued on the implants with 35 Ncm. Abut-
ment-level impression copings were con-
nected to the implant abutment platforms and
the impression procedures were carried out as
described above.
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Papaspyridakos et al
Digital versus conventional implant impressions
Fabrication of casts from conventional
impressions
Standardized pouring techniques were used
for the fabrication of all casts. After connec-
tion of the implant analogs to the impression
copings, low expansion (0.09%) type IV die
stone (Silky Rock; Whipmix Corp, Louisville,
KY, USA) was mixed. First, the stone was
mixed manually with distilled water for 15 s
to aid the incorporation of the water and
then under vacuum (Vacuum spatulator;
Whipmix Corp, Louisville, KY, USA), and an
initial pour of stone up to the middle of the
analogs was carried out. All of the stone
mixes were vibrated before and during the
pouring. After 30 min, the second pour of
vacuum-mixed die stone was carried out.
This double pouring technique minimizes
the volumetric expansion of the stone and
has been shown to lead in more accurate die
casts (Del’Acqua et al. 2008; Papaspyridakos
et al. 2012). All impressions were poured
after 2 h following impression taking to sim-
ulate a reasonable clinical scenario. The
stone casts were allowed to set for 1 h, as per
manufacturer’s recommendation, before sepa-
ration from the impressions. Subsequently,
they were trimmed and finished. All casts
were stored at room temperature for 1 week
before the measurements. The exact same
double pouring technique was used for the
fabrication of casts from all the conventional
impressions. All impression and pouring pro-
cedures were carried out by the same clini-
cian (Fig. 2a,b).
(a)
(b)
Fig. 2. (a) Splinted implant-level impression. (b) Non-
splinted implant-level impression.
Digital implant impression procedures
Following the manufacturer’s protocol, 10
repeated digital impressions were taken with
a digital intraoral scanner (TRIOS; 3shape,
Denmark) at implant level. This digital intra-
oral scanner uses confocal optical imaging
technology to generate digital point cloud
surfaces that can be exported as STL datasets
and is used for both partial and complete
arch intraoral scans.
Polymer implant impression scan bodies
(Scan bodies RC; Straumann) were connected
to the implants on the master cast (control)
by hand tightening. The digital implant
impression was gradually captured by scan-
ning the master cast and implant scan bodies
with the scanner’s handheld wand without
spraying powder. The digitally acquired vol-
umes could be viewed on the touch screen
during scanning, allowing direct visual feed-
back to make sure no parts were missing.
After the acquisition of ten repeated digital
impressions, the digital volumes were
exported as STL files for comparison.
Test groups – implant impression techniques
Thus, five test groups of casts were formed,
and each group was compared with the con-
trol cast as follows:
• GROUP I (n = 10): Stone casts generated
from the splinted coping impression tech-
nique at the implant level (internal con-
nection)
• GROUP II (n = 10): Stone casts generated
from the non-splinted coping impression
technique at the implant level (internal
connection)
• GROUP III (n = 10): Digital casts gener-
ated from the digital impression tech-
nique at the implant level (internal
connection)
• GROUP IV (n = 10): Stone casts gener-
ated from the splinted coping impression
technique at the abutment level (external
connection)
• GROUP V (n = 10): Stone casts generated
from the non-splinted coping impression
technique at the abutment level (external
connection)
Digitization of the stone casts
The four test groups of stone casts were digi-
tized for comparison with a high-resolution
extraoral scanner at 6-lm precision scanner
(IScan D103i; Imetric) as described in previ-
ous publications (Bergin et al. 2013; Ender &
Mehl 2013). At first, titanium scan bodies
(Scan markers; Dentwise, Leuven, Belgium)
were placed on the first test cast and digital
3 | Clin. Oral Impl. Res. 0, 2015 / 1–8
Papaspyridakos et al
Digital versus conventional implant impressions
scanning was performed. Then, the scan
bodies were removed and were placed on the
second cast for scanning. The same scanning
procedures were carried out for all 10 casts of
all four test groups. For all digital scans, the
same scan bodies were moved from their
mandibular corresponding position in cast 1
to cast 10 of each group to eliminate the
effect of scan bodies. An operator blinded to
the type of casts performed all scanning pro-
cedures. The STL digital files were saved.
Accuracy assessment with digital scanning and
superimposition
One digital scan of the master cast at the
implant level and one scan at the abutment
level, with the same high-resolution extraoral
scanner at 6-lm precision (IScan D103i; Imet-
ric), were used as control (golden reference)
and compared with the scans of the casts of
the five test groups produced by digital and
conventional impression techniques. To cap-
ture the 3D orientation of the implants in each
cast and their 3D discrepancies, the digital
volumes from the 5 test groups were registered
using a surface-based registration algorithm.
The 3D deviations were then calculated with
superimposition software (Mimics; Material-
ise, Leuven, Belgium) for data comparison.
The study’s workflow is shown in Fig. 3.
The terms “trueness” and “precision” rep-
resent different measures of accuracy (Ender
& Mehl 2013). Trueness is defined as the
comparison between a control dataset and a
test dataset. The measured deviations
between the control dataset and the test data-
set determine the accuracy of a scanner. Pre-
cision is defined as a comparison between
different datasets obtained using the same
Fig. 3. Study workflow.
4 | Clin. Oral Impl. Res. 0, 2015 / 1–8
digital scanner. Such an examination pro-
vides information about the repeatability of a
scanner. The digital scanning STL datasets
from all casts were imported in the computer
with dedicated software and superimposed
with the dataset from the control cast,
respectively. The 2nd parallel implant was
used as reference to superimpose the different
scanning datasets with the aid of the com-
puter software. The cumulative 3D deviation
was calculated, using the mathematical equa-
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
tion 3D ¼ x2 þ y2 þ z2 (Papaspyridakos
et al. 2012). Comparisons of all 3D deviations
were made between each of the test groups
and the control (trueness).
Statistical analysis
All statistical analyses were performed with
SAS 9.3 statistical software (SAS Institute,
Cary, NC, USA). For estimating the 3D devi-
ations of the test casts, the analysis of vari-
ance (ANOVA) and Scheffe’s post hoc test
were used to compare the differences
between Group I, Group II, and Group III,
and Student’s t-test was used for testing the
difference between Group IV and Group V.
For estimating the difference of each
implant between each of the test groups, the
nonparametric statistical methods were used.
Kruskal–Wallis test was used for comparing
the differences between Group I, Group II,
and Group III for each implant, and the dif-
ference between each test group was further
examined by Dunn’s post hoc analysis. For
testing the difference between Group IV and
Group V, Wilcoxon’s rank-sum test was used
for each implant. The level of statistical sig-
nificance was set at P < 0.0001.
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Results
The absolute values of 3D deviations from
the control cast were calculated and dis-
played for each test group in Tables 1 and 2.
The splinted, implant-level impressions
(Group I) showed median 3D (global) devia-
tions of 6 lm for implant 1 (10°), 9 lm for
implant 3 (parallel), 5 lm for implant 4 (par-
allel), and 13 lm for implant 5 (15°). The
non-splinted, implant-level impressions
(Group II) showed median 3D deviations of
22 lm for implant 1 (10°), 13 lm for implant
3 (parallel), 13 lm for implant 4 (parallel),
and 132 lm for implant 5 (15°). The digital
impressions (Group III) showed median global
deviations of 23 lm for implant 1 (10°),
15 lm for implant 3 (parallel), 8 lm for
implant 4 (parallel), and 29 lm for implant 5
(15°).
The splinted, abutment-level impressions
(Group IV) showed median 3D deviations of
33 lm for implant 1 (10°), 14 lm for implant
3 (parallel), 12 lm for implant 4 (parallel),
and 9 lm for implant 5 (15°). The non-
splinted, abutment-level impressions (Group
V) showed median 3D deviations of 15 lm
for implant 1 (10°), 1 lm for implant 3 (paral-
lel), 7 lm for implant 4 (parallel), and 10 lm
for implant 5 (15°). The overall mean values
of 3D deviation were 8, 45, 19, 17, and 8 lm
for Groups I to V, respectively.
Significant 3D deviations (P < 0.0001) were
found between Group II (non-splinted,
implant level) and control. No significant dif-
ference was found between Groups I
(splinted, implant level), III (digital, implant
level), IV (splinted, abutment level), and V
(non-splinted, abutment level) compared with
the control. As a qualitative analysis, the 3D
deviations between the test casts and the
control cast were illustrated in a color-coded
gradient (Fig. 4). When the color-difference
maps of the superimposed scans were pre-
dominantly green this indicated indicating an
exact fit between scans and the reference
model; however, red or blue color, indicating
positive or negative discrepancies, indicates
discrepancies on the fit. The box plots for the
median of individual 3D implant deviations
are shown in Figs 5 and 6.
Discussion
An accurate implant impression is necessary,
to generate an accurate definitive cast which
is the milestone for the fabrication of an
accurately fitting prosthesis. To overcome
some of the limitations with conventional
 Papaspyridakos et al
Digital versus conventional implant impressions

Table 1. Measurement and comparison of three-dimensional (3D) deviations (in lm) for Groups I, II, III
n Group I (a) Group II (b) Group III (c) P-value† Post hoc test
Overall, Median (IQR) 40 7.42 (5.28–10.88) 17.65 (13.19–76.49) 19.38 (11.54–26.21) <0.0001 (a) vs. (b)
(a) vs. (c)
Implant 1, Median (IQR) 10 5.79 (5.69–5.94) 21.89 (21.84–21.98) 23.39 (23.27–23.47) <0.0001 (a) vs. (c)
Implant 2, Median (IQR) 10 N/A N/A N/A N/A
Implant 3, Median (IQR) 10 9.16 (8.99–9.28) 13.00 (12.84–13.21) 15.27 (15.18–15.53) <0.0001 (a) vs. (c)
Implant 4, Median (IQR) 10 4.70 (4.54–4.81) 13.39 (12.97–13.46) 7.60 (7.54–7.67) <0.0001 (a) vs. (b)
Implant 5, Median (IQR) 10 12.52 (12.44–12.67) 131.75 (131.6–132.1) 29.02 (28.78–29.15) <0.0001 (a) vs. (b)
P-value <0.0001 <0.0001 <0.0001
Post hoc test Implant 4 vs. 5 Implant 4 vs. 5 Implant 4 vs. 5

N/A: implant 2 was used as a reference for superimposition.

IQR: interquartile range.
Statistical significance, P<0.0001.
†The P-value is from the Kruskal–Wallis test for each Implant.
Dunn’s post hoc tests revealed a significant difference between groups.

Table 2. Measurement and comparison of three-dimensional (3D) deviations (in lm) for Groups IV the accuracy of digital vs conventional
and V
implant impressions for the completely eden-
n Group IV Group V P-value†
tulous patient. The null hypothesis was cor-
Overall, Median (IQR) 40 13.05 (10.46–23.67) 8.23 (4.01–12.13) <0.0001
roborated. The results of this study indicate
Implant 1, Median (IQR) 10 33.10 (32.93–33.24) 14.59 (14.52–14.76) 0.0002
Implant 2, Median (IQR) 10 N/A N/A that the digital impressions had similar accu-
Implant 3, Median (IQR) 10 14.31 (13.98–14.49) 1.27 (1.19–1.37) 0.0002 racy when compared with the conventional
Implant 4, Median (IQR) 10 12.04 (11.86–12.13) 6.91 (6.69–6.96) 0.0002 impressions. Prior to the accuracy compari-
Implant 5, Median (IQR) 10 8.86 (8.81–9.01) 9.63 (9.37–9.78) 0.0002
P-value <0.0001 <0.0001
son, the casts generated from conventional
Implant 4 vs. 5 Implant 4 vs. 5 impressions had been digitized, similarly
described in previous publications to be com-
N/A: implant 2 was used as a reference for superimposition.
IQR: interquartile range. pared with the digital casts (Bergin et al.
Statistical significance, P < 0.0001. 2013; Guth et al. 2013; Kim et al. 2013).
†The P-value is from the Wilcoxon’s rank-sum test. Digital scanning and dedicated software for
superimposition of the resultant STL datasets
represent an efficient technique to measure
and compare the trueness (accuracy) at the
microscopic level (Ender & Mehl 2013). True-
ness is defined as the proximity of the abso-
lute values of the 3D deviations of each test
dataset in relation to the control dataset
(Ender & Mehl 2013; Patzelt et al. 2014). For
the present study, one parallel implant (the
second in the middle) was used during the
superimposition procedures to assess the
accuracy of implant impressions (Akyalcin
et al. 2013; Guth et al. 2013; Papaspyridakos
& Lal 2013; Schaefer et al. 2014). The super-
imposition of STL datasets by best-fit algo-
rithm has been one of the most common
methodologies to investigate the accuracy
(Guth et al. 2013). Other superimposition
techniques include the “least squares
method” and the “zero method” (Jemt &
Hjalmarsson 2012; Gimenez et al. 2015). The
accuracy outcomes may be affected by the
digital scanner, the choice of digitization
method, the alignment methodology, and the
distribution and number of surface data
Fig. 4. Color-coded gradient from Group III (digital).
points (Papaspyridakos et al. 2014a).
The findings of the present comparative
impression techniques, intraoral digital scan- in terms of accuracy of this technology study corroborated the null hypothesis and
ning was developed. Digital implant impres- (Christensen 2009). show that the accuracy of digital impressions
sions are currently gaining popularity; To the authors’ knowledge, the present in is the same as that of conventional impres-
however, limited scientific data are available vitro study is the first to directly compare sions. In regard to the implant level, the digi-

© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd 5 | Clin. Oral Impl. Res. 0, 2015 / 1–8
Papaspyridakos et al
Digital versus conventional implant impressions
Fig. 5. Implant-level groups: Absolute values of
3-D deviations from the test casts to the reference cast
(in lm).
Fig. 6. Abutment-level groups: Absolute values of
3-D deviations from the test casts to the reference cast
(in lm).
tal impressions of five mandibular implants
resulted in similar accuracy to the splinted,
implant-level impressions and both tech-
niques were superior to the non-splinted,
implant-level impression technique. The 3D
implant deviations with the non-splinted
impressions (Group II) had statistically signif-
icant differences compared with the control.
The most distal implant had a mean devia-
tion of 132 lm from the control, compared
with 13 and 29 lm for the splinted (Group I)
and the digital impressions (Group III),
respectively. These findings are in agreement
with the majority of in vitro and clinical
studies for complete arch implant impres-
sions (Papaspyridakos et al. 2011, 2014a;
Stimmelmayr et al. 2012b).
In regard to the abutment level, the 3D devi-
ations for both splinted and non-splinted
impressions were similar and did not have sta-
tistically significant differences from the con-
trol in the present study. This is in agreement
with previous in vitro studies by Kim et al.
and Del’Acqua et al., regarding abutment-
level impressions (Kim et al. 2006; Del’Acqua
et al. 2008). However, other recent in vitro
studies have shown different results from the
present study and have favoured the splinted
6 | Clin. Oral Impl. Res. 0, 2015 / 1–8
technique at the abutment level (Vigolo et al.
2003; Naconecy et al. 2004; Del’Acqua et al.
2010; Avila et al. 2012). The different method-
ologies of accuracy measurements may have
contributed to this discrepancy.
The geometry of the implant connection is
an important factor that contributes in main-
taining the stability of the implant–prosthe-
sis interface. In regard to internal connection,
the impression coping has an intimate fit
with the implant which may make with-
drawal of the impression more difficult and,
therefore, may generate a higher degree of
distortion. The connection type (internal con-
nection vs abutment level) seems to affect
accuracy because abutment-level impressions
had no statistically significant differences
from the control, whereas differences were
identified for the implant-level, non-splinted
impressions from internal connection
implants.
The accuracy of digital implant impres-
sions was not affected by the implant angula-
tion of 10° and 15° for completely edentulous
patients in the present study. This corrobo-
rated the findings of Gimenez et al. with
angulations of up to 30°, where the authors
reported that implant angulation did not
affect the accuracy in a statistically signifi-
cant rate when the blue light LAVA C.O.S.
scanner (active wavefront sampling technol-
ogy) was used (Gimenez et al. 2015). Addi-
tionally, in a duplicate study by the same
group with the identical scenario of angu-
lated implants, it was reported that implant
angulation of up to 30° did not affect statisti-
cally significantly the accuracy of digital
impressions when the red light iTero scanner
(Align Technology Inc, San Jose, CA, USA)
with parallel confocal imaging was used
(Gimenez et al. 2014). The common denomi-
nator in both studies was that the operator
experience may play a role in the accuracy of
digital impressions and that a learning curve
exists before the clinician gets skillful with
the digital impression scanners.
Only a few scientific studies have been
published regarding the accuracy of digital
impression systems for complete arch scans
with prepared teeth for tooth-supported pros-
theses. Two studies were carried out by the
same group of investigators using two digital
scanners (CEREC Bluecam & Lava C.O.S.).
The complete arch scan involved only three
prepared teeth (2 for crowns & 1 for inlay)
(Ender & Mehl 2011, 2013). The reported
accuracy (trueness) values for the CEREC
Bluecam and the Lava C.O.S. were 49  14.2
and 40.3  14.1 lm, respectively. In contrast,
accuracy values of conventional complete
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
arch impressions using polyether showed
comparable values of 55  21.8 and
61.3  21.8 lm, respectively (Ender & Mehl
2011). Another study conducted by Patzelt
et al. scanned a control cast with 14 prepared
teeth, using four different commercially
available digital scanners (CEREC Bluecam,
Lava C.O.S., iTero, Zfx IntraScan). They
showed that three of them produced compa-
rably accurate impressions (Patzelt et al.
2014). Mean accuracy values ranged from 38
to 332.9 lm, while one scanner demonstrated
unacceptable accuracy. However, it must be
mentioned that the Lava C.O.S. scanner is
not in the market anymore and has been
replaced by 3M True Definition scanner. On
the other hand, the CEREC Bluecam is still
available, but a newer improved version in
the form of CEREC Omnicam is now com-
mercially available (Yuzbasioglu et al. 2014).
A recent study by Kim et al. used the iTero
scanner with parallel confocal imaging (Align
Technology Inc) for complete arch impres-
sions of three prepared teeth and compared
the accuracy with conventional impressions
(Kim et al. 2013). The results showed that
digital impressions achieved comparable
accuracy with the conventional ones.
In terms of digital impressions for implant-
supported prostheses, there is a paucity of
scientific data limited to case reports with
single-implant crowns (Lin et al. 2013; Joda
& Bragger 2014; Wismeijer et al. 2014). A
recent in vitro study by Lee et al. investi-
gated the accuracy of digital versus closed-
tray impressions for a single-implant scenario
(Lee et al. 2014). They reported that the digi-
tal impressions had comparable accuracy
with the conventional ones. In regard to
edentulous jaws and complete arch implant
impressions, two recent in vitro studies by
Gimenez et al. used an edentulous maxilla
with six angulated implants and two types of
scanners (Lava C.O.S. and iTero, respectively)
to assess the accuracy of the digital impres-
sion (Gimenez et al. 2015; Gimenez et al.
2014). The results showed accuracy better
than 45 and 32 lm in the horizontal plane,
respectively, but there were no control
groups. Besides the present study, there are
no other studies yet comparing the accuracy
of digital vs conventional implant impres-
sions for completely edentulous patients.
A clinical study on implant impressions
for two implant-supported mandibular over-
dentures reported that the accuracy of digital
impressions with iTero was inferior to con-
ventional ones and should not be recom-
mended clinically until improvements are
made (Andriessen et al. 2014). A complete

digital workflow in implant dentistry will
include the following steps: (i) cone beam
computed tomography (CBCT) radiographic
examination followed by implant surgical
planning with virtual planning software with
or without fabrication of surgical template
for guided surgery, (ii) digital impression after
implant osseointegration, and (iii) CAD/
CAM fabrication of the implant prosthesis
(Joda & Bragger 2014). Digital impressions
include direct intraoral scanning or indirect
digitization of casts generated from conven-
tional implant impressions. The resulting
STL file form the digital impression enters
the production chain and serves as the data
for the CAD and subsequent CAM applica-
tions in a virtual “working cast-free” process.
If needed, a physical cast can be fabricated by
rapid prototyping (stereolithography, 3D
printing, or milling) from the intraoral digital
impression dataset. So far, only two clinical
reports have elaborated on the digital work-
flow for fabrication of complete arch implant
prosthesis from impression to delivery (Mo-
reno et al. 2013; Lin et al. 2014).
Previous studies have used other technol-
ogy including coordinate measuring machine
(CMM) and computed tomography for the
scanning of the master cast (control) to use
that dataset file as the golden reference
(Ender & Mehl 2013; Guth et al. 2013). For
the present study, an extraoral scanner (IScan
D103i; Imetric) with 6 lm precision was used
for all digitization procedures, which may be
seen as a limitation because other studies
have used the CMM with a repeatability of
1 lm. The TRIOS scanner for intraoral
impressions has shown acceptable clinical
References
Aguilar, M.L., Elias, A., Vizcarrondo, C.E. & Psoter,
W.J. (2010) Analysis of three-dimensional distor-
tion of two impression materials in the transfer
of dental implants. Journal of Prosthetic Den-
tistry 103: 202–209.
Akalin, Z.F., Ozkan, Y.K. & Ekerim, A. (2013) Effects
of implant angulation, impression material, and
variation in arch curvature width on implant
transfer model accuracy. International Journal of
Oral Maxillofacial Implants 28: 149–157.
Akyalcin, S., Cozad, B.E., English, J.D., Colville,
C.D. & Laman, S. (2013) Diagnostic accuracy of
impression-free digital models. American Journal
of Orthodontics and Dentofacial Orthopedics
144: 916–922.
Andriessen, F.S., Rijkens, D.R., van der Meer, W.J.
& Wismeijer, D.W. (2014) Applicability and accu-
racy of an intraoral scanner for scanning multiple
implants in edentulous mandibles: a pilot study.
Journal of Prosthetic Dentistry 111: 186–194.
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Papaspyridakos et al
Digital versus conventional implant impressions
accuracy for single crowns (Schaefer et al.
2014). Another limitation of the present
study is that only one implant system was
used at both implant and abutment level.
Further studies should be carried out with
different implant systems and scanners as
well before definitive clinical recommenda-
tions can be made for treatment of com-
pletely edentulous patients.
Digital implant dentistry is gaining
increasing popularity and is showcasing good
potential; however further studies are needed
to assess and compare the clinical accuracy
of digital versus conventional implant
impression techniques for both partially and
completely edentulous patients. Additionally,
the complete digital workflow from planning
to definitive rehabilitation should be assessed
and compared with the conventional one in
terms of time efficiency, learning curve,
accuracy, and economical aspects. In clinical
practice, combined utilization of both the
digital and the conventional approach may
present with additional advantages specific to
the treatment of each case.
Conclusions
Under the limitations of the present in vitro
study, the following conclusions may be
drawn:
• The accuracy of digital impressions was
not different than the implant-level,
splinted impressions for completely eden-
tulous patients and both more accurate
than the implant-level, non-splinted
impressions.
Avila, E.D., Moraes, F.D., Castanharo, S.M.,
Del’Acqua, M.A. & Mollo Junior, F.A. (2012)
Effect of splinting in accuracy of two implant
impression techniques. Journal of Oral Implan-
tology [epub ahead of print].
Bergin, J.M., Rubenstein, J.E., Mancl, L., Brudvik,
J.S. & Raigrodski, A.J. (2013) An in vitro compari-
son of photogrammetric and conventional com-
plete-arch implant impression techniques.
Journal of Prosthetic Dentistry 110: 243–251.
Christensen, G.J. (2009) Impressions are changing:
deciding on conventional, digital or digital plus
in-office milling. Journal of the American Dental
Association 140: 1301–1304.
Del’Acqua, M.A., Arioli-Filho, J.N., Compagnoni,
M.A. & Molo Fde, A., Jr. (2008) Accuracy of
impression and pouring techniques for an
implant-supported prosthesis. International
Journal of Oral Maxillofacial Implants 23: 226–
236.
• The implant-level, splinted impressions
were more accurate than the non-splinted
conventional impressions for completely
edentulous patients.
• The accuracy of abutment-level, splinted
impressions was not different than the
non-splinted impressions for completely
edentulous patients.
• The accuracy of implant impressions is
not affected by the implant angulation up
to 15° for completely edentulous patients.
The connection type seems to affect accu-
racy because abutment-level impressions
had no statistically significant differences
from the control, whereas differences
were identified for the implant-level, non-
splinted impressions.
Acknowledgements
The present study was funded by the ITI
Foundation, Basel, Switzerland (Research
Grant No. 753-2011). The authors wish to
express their gratitude to Chung-Han Ho,
Chi Mei Medical Center, Tainan, Taiwan, for
the statistical analysis and Dr Margarit Kha-
chatryan, Catholic University of Leuven,
Leuven, Belgium, for the digital scanning pro-
cedures. The authors do not have any finan-
cial interest in the companies whose
materials are included in this article. The
present study was conducted in partial fulfill-
ment of the requirements for the PhD degree
of Dr Papaspyridakos.
Del’Acqua, M.A., Chavez, A.M., Compagnoni,
M.A. & Molo Fde, A., Jr. (2010) Accuracy of
impression techniques for an implant-supported
prosthesis. International Journal of Oral Maxillo-
facial Implants 25: 715–721.
Ender, A. & Mehl, A. (2011) Full arch scans: con-
ventional versus digital impressions–an in-vitro
study. International Journal of Computerized
Dentistry 14: 11–21.
Ender, A. & Mehl, A. (2013) Accuracy of complete-
arch dental impressions: a new method of
measuring trueness and precision. Journal of
Prosthetic Dentistry 109: 121–128.
Gimenez, B., Ozcan, M., Martinez-Rus, F. & Pra-
dies, G. (2014) Accuracy of a digital impression
system based on active wavefront sampling tech-
nology for implants considering operator experi-
ence, implant angulation, and depth.
International Journal of Oral Maxillofacial
Implants 29: 853–862.
7 | Clin. Oral Impl. Res. 0, 2015 / 1–8
Papaspyridakos et al
Digital versus conventional implant impressions
Gimenez, B., Ozcan, M., Martinez-Rus, F. & Pra-
dies, G. (2015) Accuracy of a digital impression
system based on parallel confocal laser technol-
ogy for implants with consideration of operator
experience, implant angulation, and depth. Clini-
cal Implant Dentistry and Related Research 17
(Suppl 1): e54–e64.
Guth, J.F., Keul, C., Stimmelmayr, M., Beuer, F. &
Edelhoff, D. (2013) Accuracy of digital models
obtained by direct and indirect data capturing.
Clinical Oral Investigations 17: 1201–1208.
Jemt, T. & Hjalmarsson, L. (2012) In vitro measure-
ments of precision of fit of implant-supported
frameworks. A comparison between “virtual” and
“physical” assessments of fit using two different
techniques of measurements. Clinical Implant
Dentistry and Related Research 14(Suppl 1):
e175–e182.
Joda, T. & Bragger, U. (2014) Complete digital
workflow for the production of implant-supported
single-unit monolithic crowns. Clinical Oral
Implants Research 25: 1304–1306.
Kim, S.Y., Kim, M.J., Han, J.S., Yeo, I.S., Lim, Y.J.
& Kwon, H.B. (2013) Accuracy of dies captured
by an intraoral digital impression system using
parallel confocal imaging. International Journal
of Prosthodontics 26: 161–163.
Kim, S., Nicholls, J.I., Han, C.H. & Lee, K.W.
(2006) Displacement of implant components
from impressions to definitive casts. Interna-
tional Journal of Oral Maxillofacial Implants
21: 747–755.
Lee, S.J., Betensky, R.A., Gianneschi, G.E. & Gall-
ucci, G.O. (2014) Accuracy of digital versus con-
ventional implant impressions. Clinical Oral
Implants Research [epub ahead of print].
Lin, W.S., Harris, B.T. & Morton, D. (2013) The use
of a scannable impression coping and digital
impression technique to fabricate a customized
anatomic abutment and zirconia restoration in
the esthetic zone. Journal of Prosthetic Dentistry
109: 187–191.
Lin, W.S., Harris, B.T., Zandinejad, A. & Morton,
D. (2014) Use of digital data acquisition and
CAD/CAM technology for the fabrication of a
fixed complete dental prosthesis on dental
implants. Journal of Prosthetic Dentistry 111: 1–
5.
Ma, T., Nicholls, J.I. & Rubenstein, J.E. (1997) Tol-
erance measurements of various implant compo-
8 | Clin. Oral Impl. Res. 0, 2015 / 1–8
nents. International Journal of Oral Maxillofacial
Implants 12: 371–375.
Moreno, A., Gimenez, B., Ozcan, M. & Pradies, G.
(2013) A clinical protocol for intraoral digital
impression of screw-retained CAD/CAM frame-
work on multiple implants based on wavefront
sampling technology. Implant Dentistry 22: 320–
325.
Mpikos, P., Kafantaris, N., Tortopidis, D., Galanis,
C., Kaisarlis, G. & Koidis, P. (2012) The effect of
impression technique and implant angulation on
the impression accuracy of external- and internal-
connection implants. International Journal of
Oral Maxillofacial Implants 27: 1422–1428.
Naconecy, M.M., Teixeira, E.R., Shinkai, R.S., Fra-
sca, L.C. & Cervieri, A. (2004) Evaluation of the
accuracy of 3 transfer techniques for implant-sup-
ported prostheses with multiple abutments. Inter-
national Journal of Oral Maxillofacial Implants
19: 192–198.
Ongul, D., Gokcen-Rohlig, B., Sermet, B. & Keskin,
H. (2012) A comparative analysis of the accuracy
of different direct impression techniques for mul-
tiple implants. Australian Dental Journal 57:
184–189.
Papaspyridakos, P., Benic, G.I., Hogsett, V.L.,
White, G.S., Lal, K. & Gallucci, G.O. (2012)
Accuracy of implant casts generated with
splinted and non-splinted impression techniques
for edentulous patients: an optical scanning
study. Clinical Oral Implants Research 23: 676–
681.
Papaspyridakos, P., Chen, C.J., Chuang, S.K. &
Weber, H.P. (2014b) Implant loading protocols for
edentulous patients with fixed prostheses: a sys-
tematic review and meta-analysis. International
Journal of Oral Maxillofacial Implants 29(Suppl):
256–270.
Papaspyridakos, P., Chen, C.J., Gallucci, G.O., Dou-
koudakis, A., Weber, H.P. & Chronopoulos, V.
(2014a) Accuracy of implant impressions for par-
tially and completely edentulous patients: a sys-
tematic review. International Journal of Oral
Maxillofacial Implants 29: 836–845.
Papaspyridakos, P. & Lal, K. (2013) Computer-
assisted design/computer-assisted manufacturing
zirconia implant fixed complete prostheses: clini-
cal results and technical complications up to 4
years of function. Clinical Oral Implants
Research 24: 659–665.
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Papaspyridakos, P., Lal, K., White, G.S., Weber,
H.P. & Gallucci, G.O. (2011) Effect of splinted
and nonsplinted impression techniques on the
accuracy of fit of fixed implant prostheses in
edentulous patients: a comparative study. Inter-
national Journal of Oral Maxillofacial Implants
26: 1267–1272.
Papaspyridakos, P., Mokti, M., Chen, C.J., Benic,
G.I., Gallucci, G.O. & Chronopoulos, V. (2014c)
Implant and prosthodontic survival rates with
implant fixed complete dental prostheses in the
edentulous mandible after at least 5 years: A sys-
tematic review. Clinical Implant Dentistry and
Related Research 16: 705–717.
Patzelt, S.B., Emmanouilidi, A., Stampf, S., Strub,
J.R. & Att, W. (2014) Accuracy of full-arch scans
using intraoral scanners. Clinical Oral Investiga-
tions 18: 1687–1694.
Schaefer, O., Decker, M., Wittstock, F., Kuepper, H.
& Guentsch, A. (2014) Impact of digital impres-
sion techniques on the adaption of ceramic par-
tial crowns in vitro. Journal of Dentistry 42: 677–
683.
Stimmelmayr, M., Erdelt, K., G€ uth, J.F., Happe, A.
& Beuer, F. (2012a) Evaluation of impression
accuracy for a four-implant mandibular model–a
digital approach. Clinical Oral Investigations 16:
1137–1142.
Stimmelmayr, M., Guth, J.F., Erdelt, K., Edelhoff,
D. & Beuer, F. (2012b) Digital evaluation of the
reproducibility of implant scanbody fit–an in vi-
tro study. Clinical Oral Investigations 16: 851–
856.
Vigolo, P., Majzoub, Z. & Cordioli, G. (2003) Evalu-
ation of the accuracy of three techniques used for
multiple implant abutment impressions. Journal
of Prosthetic Dentistry 89: 186–192.
Wismeijer, D., Mans, R., van der Meer, W.J. & Reij-
ers, S.H. (2014) Patients’ preferences when com-
paring analogue implant impressions using a
polyether impression material versus digi-
tal impressions (Intraoral Scan) of dental
implants. Clinical Oral Implants Research 25:
1113–1118.
Yuzbasioglu, E., Kurt, H., Turunc, R. & Bilir, H.
(2014) Comparison of digital and conventional
impression techniques: evaluation of
patients’ perception, treatment comfort, effec-
tiveness and clinical outcomes. BMC Oral
Health 14: 10.